Eye (2003) 17, 139–148
& 2003 Nature Publishing Group All rights reserved 0950-222X/03 $25.00
Corneal sensitivity
and ocular surface
changes following
preserved amniotic
membrane
transplantation for
nonhealing corneal
ulcers
Eye (2003) 17, 139–148. doi:10.1038/
sj.eye.6700346
Introduction
Compromised ocular surface defence caused by
malfunction of lids or tear film, nerve damage,
corneal infection and chronic inflammation may
lead to persistent epithelial defects and corneal
stromal melting. Persistent corneal epithelial
defects caused by primary or secondary ocular
surface disorders may be difficult to treat and
usually induce prolonged inflammation of the
ocular surface, loss of corneal stem cells, and
destruction of the epithelial basement
membrane, resulting in corneal scarring,
neovascularisation, and decreased vision.1,2
Conservative treatment of epithelial defects
with stromal ulceration consists of treatment
modalities such as continous pressure patching
or therapeutic soft contact lenses with
preservative-free lubricants and punctum
occlusion.3 Patients failing to respond to a
medical approach may eventually require a
surgical procedure like tarsorraphy or
penetrating keratoplasty.4,5 Lee and Tseng6
initially proposed the use of human amniotic
membrane for the treatment of epithelial defects
with stromal ulcers that remain recalcitrant to
conventional treatments. Subsequently,
successful applications of amniotic membrane
transplantation (AMT) in cicatricial eye
diseases, pterygium surgery and ocular surface
reconstruction were described.7–9 Amniotic
membrane consists of a thick basement
membrane and an avascular stroma that
contains a high concentration of basic fibroblast
growth factor, basement membrane
components, and presumed trophic factors. The
www.nature.com/eye
M Dogru, M Yildiz, M Baykara, H Özc¸etin and
H Ertürk
CLINICAL STUDY
stromal matrix of the amniotic membrane is
known to have the ability to exclude the
inflammatory cells and promote healing.6 The
aim of this study was to investigate the effect of
preserved human amniotic membrane in the
treatment of nonhealing corneal ulcers and also
investigate the timewise effects on corneal
sensitivity, tear function, and impression
cytology parameters.
Materials and methods
A total of 10 eyes of 10 patients (seven males;
three females) aged between 25 and 76 years
(mean: 54.5716.5 years) underwent preserved
amniotic membrane transplantation between
April 2000 and April 2001 for persistant corneal
epithelial defects with stromal ulceration. All
patients were initially treated with removal of
toxic topical antibiotics, lubrication, autologous
serum eye drops and with a bandage contact Department of
lens/pressure patching for at least 8 weeks. A Ophthalmology
total of 44 patients diagnosed with corneal Faculty of Medicine
Uludag University
ulcers owing to infectious keratitis, alkali injury, Bursa, Turkey
diabetic neuropathy, or dry eyes responded to
the above-mentioned conventional approach Correspondence:
during the same period. Tarsorraphy was M Dogru
suggested when these techniques were Department of
unsuccessful. The 10 patients in this series did Ophthalmology
Tokyo Dental College
not opt for a tarsorraphy procedure. None of the
Ichikawa General Hospital
patients had a history of atopy, Stevens–Johnson Sugano 5-11-13, Ichikawa
syndrome, thermal or radiation injury, or Chiba 272-8513, Japan
contact lens use for refractive purposes before Tel: þ 81 47 322 0151
amniotic membrane transplantation. Three Fax: þ 87 47 322 6786
patients had type II diabetes mellitus, which E-mail: muratodooru@
yahoo.com
was not under good control (a fasting blood
glucose level of greater than 140 mg/dl and a
Received: 20 November
glycosylated haemoglobin level of greater than 2001
7.8%) throughout this study. Patients with Accepted in revised form:
herpe simplex keratitis received oral acyclovir 19 June 2002
 Ocular surface change after AMT
M Dogru et al
140

1000 mg/day as well as 60 mg/day of prednisolone for 1
week, followed by gradual tapering of oral
corticosteroids. Patients with postinfectious keratitis
received sufficient systemic and topical antibiotics/
antifungals to eradicate the causative organisms. None of
these patients had clinical or microbiologic signs of acute
corneal infection at the time of amniotic membrane
transplantation. The 10 patients in whom the corneal
epithelial defects did not close despite all mentioned
attempts received amniotic membrane transplantation.
Two patients had herpetic and three others had
postinfectious corneal ulcers. The ulcers were because of
alkali burn in two eyes and were of neurotrophic nature
in two eyes. One other eye had a persistant epithelial
defect on the corneal graft after penetrating keratoplasty.
Clinical features of the patients are summarised in Tables
1 and 2. Routine ophthalmic examinations consisted of
best-corrected visual acuity (BCVA) measurements,
anterior segment photography, and slit-lamp
examination. Corneal thickness was evaluated using
biomicroscopy or ultrasonic pachymetry (Biomed,
France) before and after the operation at each visit.
Stromal thickness was recorded and compared with the
thickness of the same area after the operation only when
the surgical area was completely epithelialised. A
stromal ulcer with keratolysis exceeding 50% of corneal
stromal thickness was regarded as a deep ulcer in this
study. Increase in stromal thickness of the eyes at the last
follow-up as observed by careful biomicroscopy and
anterior segment photography was defined as ‘stromal
gain’. The patients underwent ocular surface
examinations including corneal sensitivity
measurements, tear film break-up time (BUT), Schirmer
test, and conjunctival impression cytology. The patients
were required not to have instilled their medications at
least 6 h prior to these examinations. All examinations
were performed by the same researcher. An informed
consent about the procedures was obtained. Tear
function and corneal sensitivity measurements were
performed before amniotic membrane transplantation,
with epithelisation, attainment of resolution of the
membrane, and at the last follow-up. Corneal sensitivity
was measured using a Cochet-Bonnet aesthesiometer.
The measurements were begun with the nylon filament
fully extended. The tip of the nylon filament was applied
perpendicular to the surface of the corneal lesion making
certain not to touch the eyelashes and was pushed until
the fibre’s first visible bending. The length of the fibre
was gradually decreased until a blink reflex was
observed. The length was recorded in units of millimetre.
A corneal sensitivity measurement of less than 50 mm
was regarded as low corneal sensitivity in this study.
The standard tear film BUT measurement was
performed. Moistened fluorescein strips were introduced
into the conjunctival sac with minimal stimulation and
were undetected by the patients. The subjects were then
instructed to blink several times for a few seconds to
ensure adequate mixing of fluorescein. The interval
between the last complete blink and the appearance of
the first corneal black spot in the stained tear film was
measured three times and the mean value of the
measurements was calculated. A BUT value of less than
10 s was considered abnormal.
For further evaluation of tears, the standard Schirmer test
with topical anaesthesia (0.4% oxybuprocaine chloride) was
performed. The standardised strips of filter paper (Alcon,
TX, USA) were placed in the lateral canthus away from the
cornea and left in place for 5 min with the eyes closed.
Readings were reported in millimetres of wetting for 5 min.
A reading of less than 5 mm was referred to as dry eye.
The impression cytology specimens were obtained
after administration of topical anaesthesia with 0.4%

Table 1 Clinical features of the patients

Case Age (year)/ Cause of Size of Duration Associated ocular Prior treatment/surgery Total limbal
sex ulcer ulcer of ulcer surface and other dysfunction
(mm) (months) problems

1 46/M HSV 44 2.5 % 90 thinning PP, AS, lubricants No

2 50/M HSV 34 4.0 % 90 thinning PP, AS, lubricants No
3 75/M Neurotrophic 63 2.5 % 75 thinning, KCS TCL, AS, lubricants, No
PO, ECCE
4 65/M Neurotrophic 33 3.0 % 40 thinning, DM, KCS TCL, AS, PO, lubricants No
5 43/F Postinfectious 55 4.0 % 50 thinning SK, TCL, lubricants No
6 47/M Postinfectious 67 3.0 % 90 thinning TCL, SK, lubricants No
7 73/M Postinfectious 77 2.5 % 75 thinning, KCS, TCL, SK, lubricants No
glaucoma, DM
8 26/M Alkali burn 54 3.0 % 90 thinning, entropion, LT, SC, SK, lubricants, Yes
trichiasis, KCS, CM TCL, LFR, LME
9 45/F Alkali burn 44 4.0 % 40 thinning, KCS, CM TCL, SK, lubricants Yes
10 75/F Failed graft 67 2.5 % 30 thinning, DM PKP (2  ), PP, lubricants No

Eye
 Ocular surface change after AMT
M Dogru et al
141

Table 2 Amniotic membrane transplantation (AMT) results

Case Method of FU Preop Postop Epithelial Epitheliali- Transparency Stromal Complication Further
AMT (months) BCVA BCVA healing sation attained gain/ management
(days) phenotype (months) vascularisation

1 G+F 14 HM 20/200 14 Cornea 6 +/ None None

2 F+G+P 12 HM 20/200 25 Cornea 4 +/+ None None
3 1st G+F 13 HM 20/200 23 Cornea 7 +/+ None PKP planned,
2nd F+G+P lubricants
4 G+P 8 HM 20/200 26 Cornea 5 +/+ None PKP planned
lubricants
5 G+P 10 20/60 20/40 16 Cornea 4 +/ None None
6 F+G+P 8 HM 20/100 16 Cornea 4 +/ None PKP+ECCE
+IOL,
lubricants
7 F+G+P 8 HM HM 30 Conjunctiva None / Progressive CP, lubricants
thinning
8 F+G+P 12 HM HM 60 Conjunctiva None / None PKP planned,
lubricants
9 G+P 12 HM HM 32 Conjunctiva None / None Lubricants,
LT+PKP
planned
10 G+P 8 HM HM 45 – None / Superinfection TCL, AS,
graft failure lubricants,
repeat PKP
planned
Note: G+F: graft+filling; G+P: graft+patch; F+G+P: filling+graft+patch; HM: hand motion; PKP: penetrating keratoplasty; ECCE: extracapsular cataract
extraction; IOL: intraocular lens; CP: conjunctival patching; LT: limbal transplant; TCL: therapeutic contact lens; AS: autologous serum; FU: follow-up;
BCVA: best-corrected visual acuity.

oxybuprocaine chloride. Impression cytology was carried
out at the initial and the final follow-up visits. Strips of
cellulose acetate filter paper (Millipore HAWP 304,
Bedford, MA, USA) that were soaked in distilled water
for a few hours and dried at room temperature were
applied on the lower nasal bulbar conjunctiva adjacent to
the corneal limbus, pressed gently by a glass rod, and
then removed. The specimens were then fixed with
formaldehyde, stained with periodic acid schiff (PAS),
dehydrated in ascending grades of ethanol and then with
xylol, and finally coverslipped. The quantitative studies
of conjunctival goblet cells and squamous metaplasia of
conjunctival epithelial cells were conducted by taking
photographs using a calibrated grid under a light
microscope at a magnification of  400. We
photographed five overlapping areas of each sample
selected at random and averaged the outcomes for a
single sample score. The goblet cell densities were
reported as cells per square millimetre with standard
deviations. The specimens were also assigned a grade of
conjunctival epithelial squamous metaplasia according to
the grading scheme of Nelson.10
because of anatomic considerations. A written consent
was obtained from the mothers to harvest the amniotic
membrane. Maternal blood was screened for antibodies
against syphilis, human immunodefficiency virus,
and hepatitis virus types B and C. The placentas were
washed free of blot clots with balanced saline solution
containing 50 mg/ml of penicillin, 50 mg/ml of
streptomycin, 100 mg/ml of neomycin, and 2.5 mg/ml of
amphotericin B. The amniotic membrane was separated
from the rest of the chorion by blunt dissection. The
membranes were then flattened and sutured with the
epithelium surface up onto nitrocellulose filter papers.
The membrane with the paper was then placed in sterile
vials containing Optisol-GS (Bausch & Lomb, CA, USA)
and glycerol at a ratio of 1 : 1 (volume/volume). The
vials were then frozen and stored at 801C. The
membranes were not released for use after a second
serologic testing for HIV and hepatitis viruses from the
donor, performed 6 months after donation, proved
negative. The membranes were defrosted by warming
the vials to room temperature for 10 min immediately
before use.

Preparation of preserved human amniotic membrane Amniotic membrane transplantation

Human placentas were obtained under sterile conditions Surgery was performed under retrobulbar anaesthesia.
from planned, uneventful caesarean sections performed The base of the ulcer was debrided with a microsponge

Eye
 Ocular surface change after AMT
M Dogru et al
142
and fine forceps, and the poorly adherant epithelium
adjacent to the edge of the ulcer was removed up to the
area where the epithelium became adherant. The
amniotic membrane was then removed from the storage
medium, peeled from the nitrocellulose filter paper,
transferred to the recipient eye, and fitted to fill up the
ulcer bed and cover the defect by trimming off the excess
edges. This stroma-side down was then secured to the
edges of the ulcer by interrupted 10/0 nylon sutures and
the suture knots were buried. The decision for
transplanting more than one layer of amniotic membrane
was made by a careful assessment of preoperative and
intraoperative ulcer depth with the aid of slit-lamp
biomicroscopy and surgical microscope. More than one
layer of amniotic membrane was used if the ulcer was
deep, and in those instances, the bottom layers were left
unsutured as a filling. The second amniotic membrane
layer was transplanted as a basement membrane
(amniotic membrane graft). Depending on the aqueous
tear status and the eyelid blinking function, a third
amniotic membrane layer was transplanted as a cover
(amniotic membrane patch). When amniotic membrane
was used as a patch, this was performed by placing the
amniotic membrane over the cornea and extending it
beyond the limbus with the basement membrane side
facing down and suturing it with 10/0 interrupted nylon
sutures over the perilimbal area.
Postoperative care
Before epithelialisation, the patients were followed
weekly and were treated with topical 0.1% prednisolone
acetate (patient 10) three times a day or nonpreservative
0.4% dexamethasone three times a day (patients 1–9) and
topical 0.3% ofloxacin eye drops twice a day. After
epithelialisation was completed, the antibiotic eye drops
were discontinued but the former was gradually tapered
off over 4 months. Patients with dry eye received topical
preservative-free artificial tears or autologous serum
drops six to eight times a day as well.
Statistical analysis
Data were processed using Stat View software (1988,
Abacus Concepts, Inc., San Diego, CA, USA). The
analysis of categorised data was performed by Fisher’s
exact probability test with the probability level set at 5%
for statistical significance.
Results
Before AMT
Seven eyes in this series had received previous ocular
surgery before amniotic membrane transplantation. One
Eye
eye had received repeat keratoplasty owing to graft
failure. Corneal stroma showed variable amounts of
opacity before amniotic membrane transplantation in all
eyes. In addition, other preexisting ocular diseases such
as glaucoma and diabetic retinopathy had threatened the
visual potential in our patients. Owing to these problems,
all eyes except one had a visual acuity equal to hand
motions before amniotic membrane transplantation.
According to the respective histories, all the ulcers had
been persistent for more than 2 months, showing
progressive thinning of the ulcer bed. All corneas had
significant corneal oedema, inflammation of the limbus
and the conjunctiva.
After AMT
Depending on the depth of stromal ulceration, the ulcers
were covered by one or more than one layer of amniotic
membrane. Amniotic membrane transplantation had to
be performed twice in one patient (case 3) because of
membrane disinsertion from eye rubbing. The denuded
ulcer surfaces healed uneventfully with amniotic
membrane transplantation in all patients except case 10
between 14 and 60 days (mean: 28.7714.3 days).
Amniotic membrane transplantation could suppress the
ocular surface inflammation as evidenced by the
decrease of conjunctival and limbal hyperaemia and
corneal oedema. The ulcers healed with corneal epithelial
phenotype in six eyes in which total or partial stromal
gain could be achieved. A decrease of stromal
vascularisation was noted in three eyes (Table 2). In case
7, the corneal surface showed further thinning with signs
of imminent perforation, which led us to perform
conjunctival patching. A corneal epithelial phenotype
could not be attained in cases 8 and 9 who had total
limbal dysfunction, keratoconjunctivitis sicca, and
chronic meibomitis. The amniotic membrane was
removed in case 10 upon detection of profuse secretion
on postoperative day 6 when the membrane remnants
and the therapeutic contact lens proved positive for
Staphylococcus aereus infection. This patient was treated
with fortified vancomycin and oflaxocin eye drops
effectively and was put on list for repeat penetrating
keratoplasty upon refusal to undergo a second AMT.
During the mean follow-up period of 10.572.4 months,
visual acuity improved in the six eyes with corneal
epithelial phenotype (Table 2). Five patients lost their
therapeutic contact lenses twice until epithelialisation.
All the membranes were completely or partially
dissolved, and the remaining stroma showed variable
amounts of opacity. Figures 1–4 show the slit and
anterior segment photographs of case 2 with herpetic
ulcer before and after amniotic membrane
transplantation and at the final visit. Figures 5–7

Figure 1 Slit-lamp photography of a case with herpetic ulcer
before AMT. Corneal sensitivity: 32.5 mm, BUT: 3 s, Schirmer
test: 8 mm.
Figure 2 Anterior segment photography of the same case
before AMT.
Figure 3 Anterior segment photography after AMT performed
as filling þ graft.
Ocular surface change after AMT
M Dogru et al
143
Figure 4 Anterior segment photography at the 12th post-
operative month. Note the stromal gain and attainment of
corneal transparency. Corneal sensitivity: 47.5 mm, BUT: 9 s,
Schirmer test: 8 mm.
Figure 5 Anterior segment photography of a case with
neurotrophic ulcer before AMT. Corneal sensitivity: 20 mm;
BUT: 1 s, Schirmer test: 3 mm.
represent the anterior segment photographs of case 3
with neurotrophic ulcer before and after the first and the
second transplantation.
Corneal sensitivity
The corneal sensitivity improved in all patients except case
10 after amniotic membrane transplantation. The mean
pretransplantation corneal sensitivity was 29.574.25 mm,
which improved to a mean value of 41.7573.85 mm at
the final follow-up. The mean corneal sensitivities at the
time of attainment of amniotic membrane resolution and
at the final visit were significantly higher than the
preoperative values (Po0.05). The change of corneal
sensitivity with amniotic membrane transplantation is
Eye
 Ocular surface change after AMT
M Dogru et al
144

50
*
*
40

30

(mm)
20

10

0
Preop With With Finalvisit
epithelialisation resolution
* Fisher's test P<0.05

Figure 8 Change of corneal sensitivity with AMT. Note the

Figure 6 Anterior segment photography after the first AMT
constant increase in the mean corneal sensitivity for all patients
performed as filling þ graft. Note the decrease of limbal and
after AMT, which showed statistical significance with resolution
conjunctival hyperaemia and corneal oedema.
of the membrane and at the final visit.

60

50
Herpes simplex
40 Neurotrophic
(mm)

30 Postinfectious
Alkali burn
20
Graft failure
10

0
Preop With With Final visit
epithelialisation resolution

Figure 9 Aetiology-specific change of corneal sensitivity with

Figure 7 Anterior segment photography after second AMT
performed as filling þ graft þ patch. Note the regression of
superior and inferior corneal stromal vascularisation and
stromal gain. Corneal sensitivity: 45 mm; BUT: 6 s; Schirmer
test: 4 mm.
AMT. Note the extent of corneal sensitivity recovery in patients
with herpetic and neurotrophic ulcers compared to the others
and the decline in the sensitivity value for the patient with graft
failure at the final visit.

12
shown in Figure 8. The extent of corneal sensitivity *
10
recovery seemed to be better in patients with herpetic and
neurotrophic ulcers (Figure 9). 8
*
seconds

6
Tear function parameters The tear film BUT improved in
all patients except case 10 after amniotic membrane 4
transplantation. The mean pretransplantation BUT
2
increased from 1.570.7 to 6.572.7 s at the final visit
(Figure 10). The mean BUT values at the time of 0
Preop With With Final visit
attainment of amniotic membrane resolution and at the epithelialisation resolution
final visit were significantly higher than the preoperative * Fisher's test P<0.05
values (Po0.05). Five eyes had aqueous defficiency-type
Figure 10 Change of BUT with AMT. Note the gradual
dry eye in this study. Schirmer test values did not seem to improvement of the mean tear film BUT for all patients after
differ significantly with amniotic membrane AMT, which showed statistical significance with resolution of
transplantation (Figure 11). the membrane and at the final visit.

Eye
 Ocular surface change after AMT
M Dogru et al
145

15 2000

1500
10

Cells/mm2
Serile r1
Before AMT
mm

1000
Serile r2
Final visit

5
500

0 0
Preop With With Final visit Neurotrophic HSV Postinfectious Alkali burn Graft failure

epithelialisation resolution
Figure 13 Aetiology-specific change of goblet cell density with
Figure 11 Change of the Schirmer test with AMT. Note that the AMT. Note the marked increase in goblet cell density in patients
mean Schirmer test values did not show significant differences with herpetic and neurotrophic ulcers after AMT compared to
between the follow-up visits. the others.

3.5

2.5

2 Before AMT

1.5 Final visit

0.5

0
Neurotrophic HSV Postinfectious Alkali burn Graft failure

Figure 12 Aetiology-specific change of squamous metaplasia

grade with AMT. Note the marked improvement in the Figure 14 Impression cytology from subject with herpetic ulcer
impression cytology squamous metaplasia grade in patients before AMT: a few PAS-positive oval-plum goblet cells and
with herpetic and neurotrophic ulcers with AMT compared to sheets of small round nonsecretory epithelial cells. SM grade:
the others. 1.3370.57, goblet cell density: 5457182 cells/mm2.

Impression cytology parameters Specimens showed

variable degrees of goblet cell expression, conjunctival
epithelial squamous metaplasia, and mucin pick-up.
Except for case 10, goblet cell density and the squamous
metaplasia grade were seen to have improved at
the final visit. The mean preoperative and final goblet
cell densities were 2967250 and 6487445 cells/mm2,
respectively. The mean pretransplantation
squamous metaplasia grade improved from 1.9270.88 to
0.9271.08 at the final examination. These differences
were significant (Po0.05).
The extent of improvement in squamous metaplasia
grade and goblet cell density seemed to be better in
patients with herpetic and neurotrophic ulcers (Figures
12 and 13). Figures 14 and 15 show the impression
Figure 15 Impression cytology specimen of the same patient
cytology specimens of case 2, while Figures 16 and 17 after AMT (final visit). Note the increase of goblet cells and
represent the cytologic changes before amniotic improvement of SM grade. SM grade: 0.6670.52, goblet cell
membrane transplantation and at the last examination. density: 12727180 cells/mm2.

Eye
 Ocular surface change after AMT
M Dogru et al
146
Figure 16 Impression cytology specimen from patient with
neurotrophic ulcer before AMT. Note the loss of cellular
cohesion, nuclear pycknosis and increased SM grade. SM grade:
2.3370.57, goblet cell density: 1207104 cells/mm2.
Discussion
Previous studies reported that the amniotic membrane
has unique properties, including antibacterial, wound-
protecting, epithelialisation-promoting, and fibrosis-
suppressing effects owing to the presence of several
proteinase inhibitors, such as a1-antichymotrypsin,
a2-macroglobulin, a1-antitrypsin, a2-antiplasmin, and
inter-a1-trypsin inhibitor, as well as tissue inhibitors of
MMP-1 and MMP-2 (which inhibit destruction of stromal
collagen by MMP-2 and MMP-9), cystatin E (an analogue
of cysteine proteinase inhibitor) with complementary
antiviral properties, and also heat shock proteins with
cytoprotective functions. In addition, amniotic
membrane consists of a thick basement membrane and
an avascular stroma with a high concentration of basic
fibroblast growth factor, basement membrane
components, and presumed trophic factors, which
collectively provide benefits to augment epithelial and
stromal wound healing.11–20 We expected all these unique
properties of the amniotic membrane to improve the
epithelialisation in our patients. In addition to taking
advantage of these properties, we also took advantage of
the amniotic membrane to supplement the collagen by
using it as a filling in cases with deep stromal keratolysis.
Moreover, we made use of the amniotic membrane as a
patch in some cases to achieve constant protection and
wetting of the ulcerated area to minimise surface
exposure. We believed that a combination of collagen
layer supplementation, augmentation of corneal surface
wetting, basement membrane reconstruction, and
promotion of epithelialisation and wound healing was
required to treat severe nonhealing ulcers. In our
experience, AMT could effectively heal the ulcers similar
to other reports.21–26 The healing was with corneal
Eye
Figure 17 Impression cytology specimen from the same patient
with neurotrophic ulcer after AMT (final visit). Note the increase
of goblet cells and improvement of SM grade. SM grade:
0.6671.15, goblet cell density: 8487277 cells/mm2.
epithelial phenotype in six eyes. Three eyes healed by
conjunctival epithelial phenotype which can be
explained by the difference in the cause of the epithelial
defects, associated problems and the previous
treatments, and thus by the extent of initial damage to
the ocular surface. Failure of attainment of corneal
epithelial phenotype of healing in case 9 reconfirmed the
common belief that AMT alone is insufficient in cases
with total limbal dysfunction.27 The initial and consistent
finding in all patients was a reduction in the ocular
inflammation in the early phase after AMT as evidenced
by a decrease of conjunctival and limbal hyperaemia, and
corneal oedema. We observed that AMT could improve
visual acuity, probably owing to both corneal surface
restoration and improvement of corneal transparency.23
However, the visual results in our patients also suggest
that the effect of AMT on visual rehabilitation is limited.
A donor cornea may be the most suitable material for
grafting to treat a nonhealing ulcer and to achieve visual
rehabilitation. We think that in emergencies such as
ulcers with descemetocele and impending perforation, it
may be too late by the time a donor is found, especially
when the donor cornea supply is small. In such instances,
AMT can help the ophthalmologist buy some time by
providing corneal surface restoration and stromal gain
until a suitable graft material becomes available.
We made important observations on the corneal
sensitivity changes with AMT. The pretransplantation
corneal sensitivity was very low in all patients resulting
from herpetic/infectious keratitis, diabetes, chemical
burn, and previous surgeries, which might also have
adverse effects on limbal stem cell population. We think
that once corneal hypoesthesia sets in and persists, other
sequela on the ocular surface can develop as a result of
decreased corneal nerve trophism on epithelial functions,

tear film deficiency, and exposure owing to interruption
of corneal nerve mediated reflex arcs, that is, aqueous
tear secretion and eye lid blinking.28 To our surprise, we
saw that corneal sensitivity improved with AMT in nine
eyes, which suggested that corneal nerve regeneration
took place in our patients. Patients with herpetic keratitis
and neurotrophic ulcers seemed to have benefited most
in terms of corneal nerve recovery. It was interesting to
see that none of the patients attained full corneal
sensitivity at the final visit, which suggested that the
recovery was limited and some disease process was still
there. Even more interesting was our observation that
ulcers remained healed despite the lack of full corneal
sensitivity recovery. We would like to hypothesise that
there may be a threshold corneal sensitivity level, which
can provide sufficient trophic effects for the maintenance
of epithelial milieu and functions, and that above this
level corneal epithelium does not break down. It is also
possible that corneal epithelium may be in jeopardy
under the influence of certain disease states and that it
may start to disintegrate once this threshold value is
neared or reached. Yet, this assumption of ours has to be
tested through further clinical and experimental studies.
We used only preserved human amniotic membrane in
this study. An alternative to this might be procurement
and provision of fresh membranes when required.
However, the theoretical possibility of disease
transmission and the difficulties in finding a suitable
fresh membrane donor sufficiently far in advance of
surgery to allow processing and testing, and
coordination with admission to hospital of the recipient
should be remembered. Whether the use of preserved or
fresh amniotic membranes or different storage times will
have different effects on corneal epithelial, stromal, and
nerve healing remains to be answered through controlled
and larger studies. The changes and the differences in the
levels of presumed trophic factors between fresh and
preserved membranes and the effect of such variations
on corneal wound healing awaits further studies. It was
our belief that reduced corneal sensitivity together with
the effects of the initial chemical, infectious, or the
surgical insult would have an adverse outcome on the
tear function parameters and ocular surface cytology.
Indeed, five eyes had aqueous-deficient eyes and all eyes
had BUT-deficient dry eyes. Schirmer test results did not
change with AMT. BUT reduction most probably resulted
from ocular surface irregularity and disturbed epithelial
and goblet cell production of mucin quality and/or
quantity. We observed limited BUT improvement with
AMT owing to the attainment of epithelialisation and
possible improvement of corneal and conjunctival
epithelial mucin-related functions. We cannot answer
clearly whether the tear film instability resulted from
mucin deficiency in all patients since we did not quantify
Ocular surface change after AMT
M Dogru et al
147
the mucin content of tears this time. To be able to provide
more answers, we proceeded with conjunctival
impression cytologic analysis, which provided evidence
of squamous metaplasia, decreased goblet cell density,
and slight to no mucin pick-up in all eyes before AMT.
These parameters seemed to improve with AMT, which
point to an increase in the ocular surface health, except
for patient 10 in whom the amniotic membrane had
to be removed. The extent of improvement seemed to be
better in patients with herpetic and neurotrophic ulcers
than the patients with chemical, postinfectious, or graft
ulcers, which can be explained by the severity of the
additive effects of the initial insult on the ocular surface,
other ocular problems, and previous surgeries. To our
knowledge, the only impression cytology study related
to AMT was performed by Prabhasawat and Tseng29 in
six patients with conjunctival defects following removal
of ocular surface tumours or corneal pannus caused by
aniridia, toxic epidermal necrolysis, and chemical burns
where the defects were reconstructed by preserved
human amniotic membrane transplantation. That study
showed increased goblet cell density with AMT but lack
of corneal epithelial phenotype even on avascular
corneas, supporting the concept that conjunctival
transdifferentiation does not occur in vivo and that limbal
stem cell transplantation is needed for effective corneal
surface reconstruction in patients with total limbal
deficiency.29 However, patients with severe corneal ulcers
and limbal deficiency who underwent limbal
transplantation may still present challenges to the
clinician after AMT as in case 8 because of associated lid
problems and dry eye states. We did not encounter
serious problems after AMT. Yet, we had to remove the
amniotic membrane in case 10 upon discovery of profuse
secretion where the cultures of the amniotic membrane
and the bandage contact lens tested positive for
Staphylococcus aereus infection. The amniotic membrane
was prepared under sterile conditions and preserved in
clean conditions, and the amniotic membranes cultured at
operation were negative for microorganisms. We suspect
the therapeutic contact lens as the source of infection in
this patient, who frequently lost her lensesFone of the
common problems after AMT in our patients. We would
like to take this opportunity to remind the readership of
this article that it is very important to debride the bottom
of the ulcer and the poorly adherant epithelium at the
edges to prevent postoperative infection and to aid
healing. We would also like to draw attention to an
important observation of ours in case 7 that stromal
thinning may continue after AMT. Thus, a careful
examination for progressive keratolysis after AMT is also
essential before it is too late to take further measures.
In summary, we found that AMT can be helpful
for the treatment of epithelial defects and stromal ulcers.
Eye
 Ocular surface change after AMT
M Dogru et al
148
We came to learn that cases with total limbal deficiency
or extensive ocular surface inflammation may not be
cured with AMT alone. AMT provides improvement of
corneal sensitivity, tear film stability and impression
cytology parameters. However, the entire mechanism of
the healing effect of amniotic membrane is still unknown.
A recent experimental study revealed that amniotic
membrane transplantation vs tarsorraphy in murine
HSV-1 stromal keratitis resulted in improvements in
substance-P-related corneal nerve fibres in eyes treated
with amniotic membrane, but not in eyes treated with
tarsorraphy.30 We think that prospective, controlled,
comparative studies of corneal ulcers treated by, for
instance, tarsorraphy or other modalities should be
performed to investigate whether the beneficial changes
of amniotic membrane transplantation are greater or
lesser than the changes that might have been observed by
an alternative treatment. In this study, we tried to point
to the effects of amniotic membrane transplantation on
the ocular surface together with the problematic and
unresolved issues. Thus, it is our sincere belief that
further studies should be carried out along these lines.
References
1 Tsai RJF, Tseng SCG. Effect of stromal inflammation on
the outcome of limbal transplantation for corneal surface
reconstruction. Cornea 1995; 14: 439–449.
2 Foster CS, Zelt RP, Mai-Phan T, Kenyon KR.
Immunosuppression and selective inflammatory cell
depletion. Arch Ophthalmol 1982; 100: 1820–1824.
3 Aquavella JA. Therapeutic contact lenses. In: Aquavella JA,
Rao GN (eds). Contact Lenses. Lippincott: Philadelphia,
1987, pp 140–163.
4 Pakarinen M, Tervo T, Tarkkanen A. Tarsorraphy in the
treatment of persistent corneal lesions. Acta Ophthalmol
Scand 1987; 182(Suppl): 69–73.
5 Dohlman CH, Boruchoff SA, Sullivan GL. A technique for
the repair of perforated corneal ulcers. Arch Ophthalmol
1967; 73: 471.
6 Lee SH, Tseng SCG. Amniotic membrane transplantation for
persistent epithelial defects with ulceration. Am J
Ophthalmol 1997; 123: 303–312.
7 Tsubota K, Satake Y, Ohyama M. Surgical reconstruction
of the ocular surface in advanced cicatricial pemphigoid
and Stevens Johnson Syndrome. Am J Ophthalmol 1996;
122: 38–52.
8 Prabhasawat P, Barton K, Burkett G, Tseng SCG.
Comparison of conjunctival autografts and primary closure
for pterygium excision. Ophthalmology 1997; 104: 974–985.
9 Shimazaki J, Yang HY, Tsubota K. Amniotic membrane
transplantation for ocular surface reconstruction in patients
with chemical and thermal burns. Ophthalmology 1977;
104: 2068–2076.
10 Nelson JD. Impression cytology. Cornea 1988; 7: 71–81.
11 Dua HS, Azuara-Blanco A. Amniotic membrane
transplantation. Br J Ophthalmol 1999; 83: 748–752.
12 Na BK, Hwang JH, Kim JC. Analysis of human amniotic
membrane components as proteinase inhibitors for
Eye
development of therapeutic agent for recalcitrant keratitis.
In: Masaomi T, Toshio H, Ichiro T (eds). Trophoblast Research.
New York: University of Rochester Press, 1999, pp 453–466.
13 Riley SC, Leask R, Denison FC. Secretion of tissue inhibitors
of matrix metalloproteinase by human fetal membrane,
decidua and placenta at parturition. J Endocrinol 1999; 162:
351–359.
14 Shinozaki M, Shoda A, Shimazaki J. Detection of basic
fibroblast growth factor from amniotic membrane. Invest
Ophthalmol Vis Sci 1995; 36: 131–138.
15 Lwebuga-Mukasa JS, Thulin G, Madri JA. An acellular
human amniotic membrane model for in vitro culture of
type II pneumocytes, the role of basement membrane in cell
morphology and function. J Cell Physiol 1984; 121: 215–225.
16 Modesti A, Scarpa S, D’Orazi G. Location of type IV and V
collagens in the stroma of human amnion. Prog Clin Biol Res
1989; 296: 459–463.
17 Behzad F, Jones CJ, Aplin JD. The role of integrin alpha 6
and 4 in hemidesmosomes of human amnion. Biochem Soc
Trans 1991; 19: 389S.
18 Kurpakus MA, Stock EI, Jones JCR. The role of basement
membrane in differential expression of keratin proteins in
epithelial cells. Dev Biol 1992; 150: 243–255.
19 Ni J, Abrahamson M, Zhang M. Cystatin E is a novel human
cysteine proteinase inhibitor with structural resemblance to
family 2 cystatins. J Biol Chem 1997; 272: 10853–10858.
20 Cheong TB, Ko YS, Kwon NS. Heat shock protein can
inhibit corneal fibroblast death from oxidative stress. Invest
Ophthalmol Vis Sci 1999; 40: S3291.
21 Kruse FE, Rohrschneider K, Völcker HE. Multiple layer
amniotic membrane transplantation for reconstruction of
deep corneal ulcers. Ophthalmology 1999; 108: 1504–1511.
22 Hanada K, Shimazaki J, Shimmura S, Tsubota K.
Multilayered amniotic membrane transplantation for severe
ulceration of the cornea and sclera. Am J Ophthalmol 2001;
131: 324–331.
23 Letko E, Stechschulte SU, Kenyon KR, Sadeq N, Romero TR,
Samson CM. Amniotic membrane inlay and overlay
grafting for corneal epithelial defects and stromal ulcers.
Arch Ophthalmol 2001; 119: 659–663.
24 Kim JS, Kim JC, Hahn TW, Park WC. Amniotic membrane
transplantation in infectious corneal ulcer. Cornea 2001; 20:
720–726.
25 Chen HJ, Pires RTF, Tseng SCG. Amniotic membrane
transplantation for severe neurotrophic ulcers. Br J
Ophthalmol 2000; 84: 826–833.
26 Gabric N, Mravicic I, Dekaris I, Karaman Z, Mitrovic S.
Human amniotic membrane in the reconstruction of the
ocular surface. Doc Ophthalmol 1999; 98: 273–283.
27 Anderson DF, Ellies P, Pires RTF, Tseng SCG. Amniotic
membrane transplantation for partial limbal stem cell
deficiency. Br J Ophthalmol 2001; 85: 567–575.
28 Tseng SCG, Tsubota K. Important concepts for treating
ocular surface and tear disorders. Am J Ophthalmol 1997;
124: 825–835.
29 Prabhasawat P, Tseng SCG. Impression cytology study of
epithelial phenotype of ocular surface reconstructed by
preserved human amniotic membrane. Arch Ophthalmol
1997; 115: 1360–1367.
30 Heilingenhaus A, Bauer D, Meller D, Steuhl KP, Tseng SCG.
Improvement of HSV-1 necrotizing keratitis with amniotic
membrane transplantation. Invest Ophthalmol Vis Sci 2001;
42: 1969–1974.